Connecting Using: sqlplus "/ as sysdba" Fails with ORA-12547 with Lower Ulimit Setting

(Doc ID 878375.1)

Last updated on DECEMBER 06, 2016

Applies to:

Oracle Database - Enterprise Edition - Version 8.1.7.0 to 10.2.0.4 [Release 8.1.7 to 10.2]
Oracle Solaris on SPARC (64-bit)
Sun Solaris SPARC (64-bit)
***Checked for relevance on 06-Aug-2016***

Symptoms

$sqlplus "/ as sysdba" fails with ORA-12547 .

You may have more than one database up and running under the same oracle home . When you try to bring up a database while connecting as "/ as sysdba" , the BEQ connection fails with ORA-12547 .

You might see significant delay after the sqlplus "/ as sysdba" is fired and before ORA/TNS-12547 occurs.

The problem still persists even after performing $relink all  . No error/warning messages in the OS logs.

sqlnet client trace may show the following error stack :

[31-JUL-2009 14:26:43:569] sntpcall:   environ[19] = "ORACLE_HOME=/oracle/product/10.2.0.1"
[31-JUL-2009 14:26:43:570] sntpcall:   environ[20] = "PWD=/export/home/ora10201"
[31-JUL-2009 14:26:43:570] sntpcall:   environ[21] = "TZ=Australia/Queensland"
[31-JUL-2009 14:27:21:150] sntpcall: only 0 bytes read
[31-JUL-2009 14:27:21:150] sntpcall: Can't read from pipe; err[1] = 32
[31-JUL-2009 14:27:21:151] sntpcall: exit
[31-JUL-2009 14:27:21:151] ntpcon: exit
[31-JUL-2009 14:27:21:152] nserror: entry
[31-JUL-2009 14:27:21:152] nserror: nsres: id=0, op=65, ns=12547, ns2=12560; nt[0]=517, nt[1]=32,
nt[2]=0; ora[0]=0, ora[1]=0, ora[2]=0
[31-JUL-2009 14:27:21:153] nsopen: unable to open transport
[31-JUL-2009 14:27:21:153] nsiocancel: entry
[31-JUL-2009 14:27:21:154] nsiofrrg: entry
[31-JUL-2009 14:27:21:154] nsbfr: entry
...........
.............
[31-JUL-2009 14:27:21:167] nscall: error exit
[31-JUL-2009 14:27:21:167] nioqper:  error from nscall
[31-JUL-2009 14:27:21:167] nioqper:    ns main err code: 12547
[31-JUL-2009 14:27:21:168] nioqper:    ns (2)  err code: 12560
[31-JUL-2009 14:27:21:168] nioqper:    nt main err code: 517
[31-JUL-2009 14:27:21:169] nioqper:    nt (2)  err code: 32
[31-JUL-2009 14:27:21:169] nioqper:    nt OS   err code: 0

Truss on sqlplus may look like :

...........
...........
4074:  execve("/oracle/product/10.2.0.1/bin/oracle", 0x10010D6F0, 0x100217D40)  argc = 2
.......
4074:    (DESCRIPTION=(LOCAL=YES)(ADDRESS=(PROTOCOL=beq)))
........
..........
4074:  shmat(9, 0x380000000, 040000)      = 0x380000000
4074:      Incurred fault #6, FLTBOUNDS  %pc = 0x100E59824
4074:        siginfo: SIGSEGV SEGV_MAPERR addr=0x000006C0
4074:      Received signal #11, SIGSEGV [caught]
4074:        siginfo: SIGSEGV SEGV_MAPERR addr=0x000006C0
4074:  lwp_sigmask(SIG_SETMASK, 0x9FBEF457, 0x0000FFF7) = 0xFFBFFEFF [0x0000FFFF]
4074:  lwp_sigmask(SIG_SETMASK, 0x9FBEF057, 0x0000FFF7) = 0xFFBFFEFF [0x0000FFFF]
4074:      Incurred fault #6, FLTBOUNDS  %pc = 0x100E5BD3C
4074:        siginfo: SIGSEGV SEGV_MAPERR addr=0x00000008

The ulimit settings may look something like :

  $ ulimit -a
    core file size        (blocks, -c) unlimited
    data seg size         (kbytes, -d) unlimited
    file size             (blocks, -f) unlimited
    open files                    (-n) 256
    pipe size          (512 bytes, -p) 10
    stack size            (kbytes, -s) 8192
    cpu time             (seconds, -t) unlimited
    max user processes            (-u) 29995
    virtual memory        (kbytes, -v) unlimited

Cause

Sign In with your My Oracle Support account

Don't have a My Oracle Support account? Click to get started

My Oracle Support provides customers with access to over a
Million Knowledge Articles and hundreds of Community platforms